Using the GNU Compiler Collection (GCC): PowerPC Function Attributes

Using the GNU Compiler Collection (GCC): PowerPC Function Attributes

6.31.23 PowerPC Function Attributes

These function attributes are supported by the PowerPC back end:

longcall

shortcall

The longcall attribute indicates that the function might be far away from the call site and require a different (more expensive) calling sequence. The shortcall attribute indicates that the function is always close enough for the shorter calling sequence to be used. These attributes override both the -mlongcall switch and the #pragma longcall setting.

See RS/6000 and PowerPC Options, for more information on whether long calls are necessary.

target (options)

As discussed in Common Function Attributes, this attribute allows specification of target-specific compilation options.

On the PowerPC, the following options are allowed:

‘altivec’
‘no-altivec’: Generate code that uses (does not use) AltiVec instructions. In 32-bit code, you cannot enable AltiVec instructions unless -mabi=altivec is used on the command line.
‘cmpb’
‘no-cmpb’: Generate code that uses (does not use) the compare bytes instruction implemented on the POWER6 processor and other processors that support the PowerPC V2.05 architecture.
‘dlmzb’
‘no-dlmzb’: Generate code that uses (does not use) the string-search ‘dlmzb’ instruction on the IBM 405, 440, 464 and 476 processors. This instruction is generated by default when targeting those processors.
‘fprnd’
‘no-fprnd’: Generate code that uses (does not use) the FP round to integer instructions implemented on the POWER5+ processor and other processors that support the PowerPC V2.03 architecture.
‘hard-dfp’
‘no-hard-dfp’: Generate code that uses (does not use) the decimal floating-point instructions implemented on some POWER processors.
‘isel’
‘no-isel’: Generate code that uses (does not use) ISEL instruction.
‘mfcrf’
‘no-mfcrf’: Generate code that uses (does not use) the move from condition register field instruction implemented on the POWER4 processor and other processors that support the PowerPC V2.01 architecture.
‘mfpgpr’
‘no-mfpgpr’: Generate code that uses (does not use) the FP move to/from general purpose register instructions implemented on the POWER6X processor and other processors that support the extended PowerPC V2.05 architecture.
‘mulhw’
‘no-mulhw’: Generate code that uses (does not use) the half-word multiply and multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors. These instructions are generated by default when targeting those processors.
‘multiple’
‘no-multiple’: Generate code that uses (does not use) the load multiple word instructions and the store multiple word instructions.
‘update’
‘no-update’: Generate code that uses (does not use) the load or store instructions that update the base register to the address of the calculated memory location.
‘popcntb’
‘no-popcntb’: Generate code that uses (does not use) the popcount and double-precision FP reciprocal estimate instruction implemented on the POWER5 processor and other processors that support the PowerPC V2.02 architecture.
‘popcntd’
‘no-popcntd’: Generate code that uses (does not use) the popcount instruction implemented on the POWER7 processor and other processors that support the PowerPC V2.06 architecture.
‘powerpc-gfxopt’
‘no-powerpc-gfxopt’: Generate code that uses (does not use) the optional PowerPC architecture instructions in the Graphics group, including floating-point select.
‘powerpc-gpopt’
‘no-powerpc-gpopt’: Generate code that uses (does not use) the optional PowerPC architecture instructions in the General Purpose group, including floating-point square root.
‘recip-precision’
‘no-recip-precision’: Assume (do not assume) that the reciprocal estimate instructions provide higher-precision estimates than is mandated by the PowerPC ABI.
‘string’
‘no-string’: Generate code that uses (does not use) the load string instructions and the store string word instructions to save multiple registers and do small block moves.
‘vsx’
‘no-vsx’: Generate code that uses (does not use) vector/scalar (VSX) instructions, and also enable the use of built-in functions that allow more direct access to the VSX instruction set. In 32-bit code, you cannot enable VSX or AltiVec instructions unless -mabi=altivec is used on the command line.
‘friz’
‘no-friz’: Generate (do not generate) the friz instruction when the -funsafe-math-optimizations option is used to optimize rounding a floating-point value to 64-bit integer and back to floating point. The friz instruction does not return the same value if the floating-point number is too large to fit in an integer.
‘avoid-indexed-addresses’
‘no-avoid-indexed-addresses’: Generate code that tries to avoid (not avoid) the use of indexed load or store instructions.
‘paired’
‘no-paired’: Generate code that uses (does not use) the generation of PAIRED simd instructions.
‘longcall’
‘no-longcall’: Generate code that assumes (does not assume) that all calls are far away so that a longer more expensive calling sequence is required.
‘cpu=CPU’: Specify the architecture to generate code for when compiling the function. If you select the target("cpu=power7") attribute when generating 32-bit code, VSX and AltiVec instructions are not generated unless you use the -mabi=altivec option on the command line.
‘tune=TUNE’: Specify the architecture to tune for when compiling the function. If you do not specify the target("tune=TUNE") attribute and you do specify the target("cpu=CPU") attribute, compilation tunes for the CPU architecture, and not the default tuning specified on the command line.

On the PowerPC, the inliner does not inline a function that has different target options than the caller, unless the callee has a subset of the target options of the caller.